Data Mapping for Adobe Analysts

This table shows the nature of the data collected and where it is stored in Adobe Analytics.

Stackchat Trackable Item Prop eVar Events to Track Description Notes / Questions
pageName Not populated
pageURL Not populated See note 1 below
Channel channel
(e.g Facebook, Web, WeChat)
Visitor Id visitorID As per ECID requirements
Conversation Id prop1 evar1 MESSAGE_SENT Fires whenever any message is sent.
MESSAGE_RECEIVED Fires whenever any message is received.
POSTBACK_RECEIVED Fires whenever a postback is received.
CAROUSEL_SENT
IMAGE_SENT
TEXT_SENT Only fires for a text message declared in CDML. Not, for example, when we send a text prompt before a quick nav.
LIST_SENT
QUICKNAV_SENT Only fires once for a quick nav, even though we send both the prompt message and the actual navs.
Language prop20
Input Method prop22 Can only be "voice" or "text" and is only relevant for the SDKs that support both input methods.
User Input Group prop2 evar2 Since User Input Groups don't have a 'name', this is tracked as the value: "${MESSAGE_THREAD_NAME}.USER_INPUT[X]" where X denotes the index of the user input in the message thread (since there could be multiple user input groups in a given thread).
START
COMPLETE Fires on successful completion of the entire user input group.
CORRECTION_MADE Fires if the user chooses to adjust a field following the confirmation prompt.
User Input prop3 evar3 Tracks the label of the user input (e.g 'WineType') - not the value passed by the user.
VALIDATION_FAILED Fires if the user provides an invalid value (for example, a number if the slot is of type 'date').
Logic Branch prop4 evar4 TRUE Fires when the branch resolves as true. Tracks the name of the logic branch.
FALSE Fires when the branch resolves as false.
Cloud Function prop5 evar5 START Tracks the name of the cloud function.
COMPLETE
ERROR Fires when there's a runtime exception in the code.
Cloud Function Error prop21 Tracks the type of error (null-pointer, execution timeout, etc.).
NLP Intent Name prop6
NLP Intent Type prop7 QA vs Nav Intent
NLP Confidence prop8 Registers as a number between 0 and 1, rounded to 2 decimal places.
NLP Confidence Range prop23
NLP Entities Extracted prop9 This tracks the names of the entities that were extracted, not the values passed by the user. Styled as a comma-separated list.
NLP user query prop10
Sentiment Score prop11
Sentiment Magnitude prop12
Action Sequence prop13 Tracks the name of the Action Sequence.
Clear Slots Action prop14 Tracks the name of the Clear Slots action (e.g erase_attraction_facilities_choice).
Set Slots Action prop15 Tracks the name of the Set Slots action.
Analytics Event Action prop16 Tracks the name you assigned to the analytics action.
Message Thread prop17 Tracks the name of the Message Thread.
Navigation prop18 Tracks the value of the flow and flow element navigated to in the form of Flow / Element.
Incoming User Message prop19 Track the value of the incoming message sent by the user.

Note 1: Flows could be considered as pages and the Flow / Flow Element concept could be considered as a pageURL, but these don't translate well, because in a conversation, your bot may jump around between a handful of Flows and Elements in response to a single user query. This results in a confusing set of data (e.g. 7 page views detected in response to a single user query), so we've opted to NOT populate the pageName and pageURL properties.

If you want to see how many times a given flow has been hit, you can see that in the Navigation report. If you want to see how many times users have queried your bot, you can see that in the User Query report.